This is the mail archive of the
gdb-prs@sources.redhat.com
mailing list for the GDB project.
Re: java/1907: gdb does not notice when a Java native library is loaded via classmap.db
- From: Robin Green <greenrd at greenrd dot org>
- To: nobody at sources dot redhat dot com
- Cc: gdb-prs at sources dot redhat dot com,
- Date: 11 Apr 2005 11:58:01 -0000
- Subject: Re: java/1907: gdb does not notice when a Java native library is loaded via classmap.db
- Reply-to: Robin Green <greenrd at greenrd dot org>
The following reply was made to PR java/1907; it has been noted by GNATS.
From: Robin Green <greenrd@greenrd.org>
To: gdb-gnats@sources.redhat.com
Cc:
Subject: Re: java/1907: gdb does not notice when a Java native library is loaded via classmap.db
Date: Mon, 11 Apr 2005 12:53:25 +0100
On Sat, Apr 09, 2005 at 11:29:54AM +0100, Andrew Haley wrote:
> > gdb does not notice when a Java native library is loaded via
> > classmap.db (the new "BC API"). (Not very surprising, since this
> > feature of gcc-java is quite new!)
>
> I am certain that you are wrong about this. I have been using this
> feature for at least a year, and if it you were right I would not have
> been able to get anything done at all.
Apologies. The real problem, I think, is that I was using gdb to attach to
another process. When I instead launched the process from gdb, it worked.
You mentioned in IRC that you thought this was a known issue, because the
implementation cannot register to receive shlib load events after the program
has already started.
However, I just searched for an existing bug along the lines of "gdb doesn't notice
libs being loaded after being attached to a running process", but I
couldn't find a PR in the database that covers this.
I guess there are actually two bugs here:
1) The limitation itself;
2) gdb fails to warn you about the limitation at runtime - and it even offers
to set a pending breakpoint for you, an offer which it cannot fulfill,
apparently!
I'll file two new bugs for these tommorow, unless advised otherwise.
--
Robin